What Is Related To What In My Data?

Andy Grogan-Kaylor

In some recent projects, I have had to get a quick idea of what scales and demographic variables in my data are related to other scales and demographic variables.

It seems to me that one of the quickest ways to do this is with the corrgram (Wright, 2021) library. I also like pairs.panels in the psych (Revelle, 2025) library, and ggpairs in the GGally (Schloerke et al., 2025) library.

Lastly, a network graph using igraph (Csardi et al., 2025; Csardi & Nepusz, 2006) requires substantially more programming, and produces a quite different looking visualization that may or may not be useful in some situations.

Call Libraries

Show the code

library(MASS) # simulate multivariate data

library(pander) # nice tables

library(corrgram) # corrgrams

library(GGally) # pairs plots

library(psych) # useful routines for psych data

library(igraph) # network analysis

Simulated Data

I simulate some data.

Key Idea

Of particular note are that the strongest correlation is between scaleA and scaleB, and that scaleA and scaleC are negatively correlated.

Show the code

set.seed(54321)

# simulate data

mu <- c(scaleA = 10, scaleB = 25, scaleC = 100) # means

Sigma <- matrix(c(1.0, 0.75, -0.3, # variance covariance
                  0.75, 1.0, 0.4,
                  -0.3, 0.4, 1.0), 
                nrow = 3, ncol = 3)

mydata <- mvrnorm(n = 500, 
                  mu = mu, 
                  Sigma = Sigma) # variance covariance

Look At The Data

Show the code

pander(head(mydata)) # nicely formatted table

scaleA	scaleB	scaleC
9.291	25.14	101.2
8.853	24.3	100.4
9.186	24.35	99.86
7.81	23.8	101.4
9.734	24.54	99.8
8.603	24.16	100.9

Show the code

skimr::skim(mydata) # descriptive statistics

Variable type: numeric

skim_variable	complete_rate	mean	sd	p0	p25	p50	p75	p100	hist
scaleA	1	10.00	1.04	7.14	9.30	9.97	10.68	13.30	▁▅▇▃▁
scaleB	1	24.94	1.00	22.12	24.27	24.92	25.62	27.97	▁▅▇▅▁
scaleC	1	99.91	1.00	97.20	99.25	99.92	100.56	102.82	▁▅▇▃▁

What Is Related To What In My Data?

Show the code

corrgram(mydata, 
         order=TRUE, 
         lower.panel = panel.pie, # lower panel is pie charts
         upper.panel = panel.cor, # upper panel is correlations
         text.panel = panel.txt, # text panel is variable
         diag.panel=panel.minmax, # minimum and maximum
         main = "Correlation of Scales in Our Data")

Show the code

# ggpairs(mydata) # simple default plot

ggpairs(mydata, # pairs plot
        lower = list(continuous = wrap("points", 
                                       color = "blue", 
                                       alpha = 0.5)),
        diag = list(continuous = wrap("densityDiag",
                                      color = "blue")),
        upper = list(continuous = wrap("cor",
                                       color = "black"))) + 
  labs(title = "Correlation of Scales in Our Data")

Show the code

pairs.panels(mydata, # pairs plot
             show.points = FALSE, # don't show data points
             main="Correlation of Scales in Our Data")

Show the code

cor_mat <- cor(mydata, use = "complete.obs")

diag(cor_mat) <- 0 # set diagonal to 0

g <- graph_from_adjacency_matrix(cor_mat, # make network graph
                                 mode = "undirected", 
                                 weighted = TRUE)

# par(mar=c(0, 0, 1, 0) + 1.5) # plot margins

absweight <- abs(E(g)$weight) # absolute value of weight

plot(g, 
     margin = 0.5,
     edge.width = absweight * 20, # absolute value
     vertex.size = 15,
     vertex.label.dist = 3.5,
     vertex.shape = "circle",
     vertex.color = "dodgerblue",
     edge.label.color = "red",
     edge.label = round(E(g)$weight, 2),
     edge.label.dist = 1, 
     # layout = coords,
     edge.curved = 0,
     main = "Correlation of Scales in Our Data")

References

Csardi, G., & Nepusz, T. (2006). The igraph software package for complex network research. InterJournal, Complex Systems, 1695. https://igraph.org

Csardi, G., Nepusz, T., Traag, V., Horvat, S., Zanini, F., Noom, D., & Müller, K. (2025). igraph: Network analysis and visualization in R. https://doi.org/10.5281/zenodo.7682609

Revelle, W. (2025). Psych: Procedures for psychological, psychometric, and personality research. Northwestern University. https://CRAN.R-project.org/package=psych

Schloerke, B., Cook, D., Larmarange, J., Briatte, F., Marbach, M., Thoen, E., Elberg, A., & Crowley, J. (2025). GGally: Extension to ’ggplot2’. https://doi.org/10.32614/CRAN.package.GGally

Wright, K. (2021). Corrgram: Plot a correlogram. https://doi.org/10.32614/CRAN.package.corrgram